US 7133425 B2
A method and apparatus is disclosed for facilitating high-speed data communication. In one embodiment a receiver is configured to include error correction coding in an interleaved environment to increase data transmission rates. In one embodiment the signal mapping scheme includes constellation shaping to reduce transmit power levels. In one embodiment 8-state Ungerboeck trellis coding is implemented with 4-dimensional PAM10 signal mapping.
1. A communication system for transmitting data over two or more channels comprising:
a de-multiplexer configured to receive a first number of data streams from a data source and output a second number of data streams, wherein the second number is greater than the first number;
one or more error correction encoders configured to receive the second number of data streams and perform encoding on the second number of data streams to create encoded data;
a multiplexer configured to process the encoded data from the one or more error correction encoders and output encoded data in a third number of data streams; and
one or more mapping modules configured to receive encoded data in the third number of data streams from the multiplexer and generate symbol values representative of encoded data in the third number of data streams, the symbol values selected from a constellation of available symbol values.
2. The system of
3. The system of
4. The system of
5. The system of
6. The system of
7. The system of
8. A method of processing data prior to transmission over two or more channels to increase the data transmission rate, the method comprising:
receiving a first data stream from a data source;
demultiplexing the first data stream into two or more second data streams;
performing error correction encoding on the two or more second data streams to create two or more encoded data streams;
multiplexing the two or more encoded data streams to create interleaved encoded data streams wherein the number of interleaved encoded data streams is less than the number of encoded data streams; and
modulating the interleaved encoded data streams to convert the interleaved encoded data streams into symbol values on two or more modulator outputs.
9. The method of
10. The method of
11. The method of
12. The method of
13. A communication system for processing at least one signal received over two or more channels comprising:
one or more equalizers configured to receive at least one signal from the two or more channels and process the at least one signal to reduce the effects of transmission through the two or more channels;
a de-multiplexer configured to receive the at least one signal from the one or more equalizers on a first number of inputs and output at least two signals on a second number of outputs, wherein the second number is greater than the first number;
one or more error correction decoders configured perform decoding on the at least two of the said second number of outputs creating at least two decoded signals;
a multiplexer configured to combine the at least two decoded signals into at least one output decoded signal; and
one or more demapping modules configured to translate the at least one output decoded signal to data.
14. The system of
15. The system of
16. The system of
17. The system of
18. The system of
19. A method for processing signals received at a first data rate comprising:
receiving two or more signals via two or more channels;
filtering the two or more signals to reduce the effects of intersymbol interference;
de-interleaving the two or more signals to create three or more signals at a second data rate;
error correction processing the three or more signals to correct errors arising during transmission to create three or more error corrected signals;
interleaving the three or more error corrected signals two or more interleaved signals at a third data rate;
demodulating the two or more interleaved signals to convert the two or more interleaved signals to data.
20. The method of
21. The method of
22. The method of
23. The method of
24. The method of
25. A system for achieving data communication between a first location and a second location comprising:
a transmitter comprising:
means for receiving a full rate data stream;
means for time de-interleaving the full rate data stream into two or more fractional rate data streams;
means for performing error correction coding on the two or more fractional rate data streams to create two or more fractional rate encoded data streams;
means for time interleaving the two or more fractional rate encoded data streams into a full rate encoded data stream;
means for mapping the full rate encoded data stream into symbol values;
means for precoding the symbol values;
means for transmitting the precoded symbol values to second location;
a receiver at the second location, configured to receive the precoded symbol values, comprising;
means for filtering the precoded symbol values to compensate for the effects of transmitting to a second location to create a full rate filtered symbol stream;
means for time de-interleaving the full rate filtered symbol values to create a fractional rate filtered symbol stream;
means for decoding the fractional rate filtered symbol stream to create fractional rate decoded symbol stream;
means for time interleaving the fractional rate decoded symbol stream to create full rate decoded symbol stream;
means for demapping the full rate decoded symbol stream to create a full rate demodulated data stream.
26. The system of
27. The system of
28. The system of
The invention relates to communication systems and in particular to a method and apparatus for data encoding and decoding.
Demand for high rate data transmission continues to increase in response to new service offerings and expanded communication network usage, such as for home and business tasks. For example, audio and video content is growing in popularity and is generally bandwidth intensive. In addition, many tasks are now commonly handled via a computer network, such as ordering business supplies, exchanging documents, and information gathering. Moreover, bandwidth demand placed upon a company's local area network is increasing. Often networks are being burdened with more users and larger and more complex software applications. These applications are bandwidth intensive and the complex software applications create larger files. Although many networks are currently at bandwidth capacity additional demands are continually being placed on these network.
While there are numerous proposed solutions to alleviate network congestion by increasing network speeds, many of these proposed solutions require migration to a different communication standard or a different communication medium than is currently in use. While different communication standards (SONET, ATM) and different communication mediums (fiber optic cable, coaxial cable) may make it possible to increase data rates, the cost associated with migration is extremely prohibitive. For example, installation of an optical network to each computer in a local area network requires significant software and hardware upgrades on each end user's platform and installation of fiber optic cable. The cabling cost alone is tremendous.
Another proposed solution is to increase the transmission rate of existing networks. One widely deployed network standard is Ethernet, which utilizes twisted pairs of copper wires as its transmission medium. Although widely deployed and inexpensive to manufacture, twisted pair copper is bandwidth limited. As a result, data signals transmitted at a high rate over the twisted pair copper, or other medium, are subject to significant levels of distortion and error rates. Upon reception, recovering the transmitted signal may be impossible due to this distortion.
While it is possible to perform signal processing or error correction coding, such prior art attempts have been insufficient to achieve desired data transmission rates with acceptable error rates. For example, some attempts to maximize the throughput and signal to noise margin of the channel through use of integrated circuit systems are simply not possible or feasible with existing semiconductor manufacturing processes.
Previous attempts to achieve these above-mentioned goals have not achieved the desired results, which is high data rate, low error rate, data communication. In general, the prior art attempts may be divided into precoded and non-precoded systems. For example, the Rao reference, (U.S. Pat. No. 6,088,827), proposes a non-precoded system that combines PAM5 with Trellis Coded Modulation at a baud rate appropriate for transmitting an aggregate of 1 Gbps over Category-5E unshielded twisted pair (UTP). However, the solution proposed in the Rao reference suffers from several shortcomings. One such shortcoming is that additional symbols must be added to the packet to terminate the trellis. This however, reduces the data transmission rate. Furthermore, control symbols, such as Start and Terminate are not detected using the proposed trellis decoder, and as a result, the control symbols do not enjoy a coding gain. In addition, the PAM-5 mapping as contemplated by the Rao reference is not balanced and thus requires an additional scrambling technique to create a DC balance on the transmit signal. For these reasons, the method and apparatus proposed by the Rao reference does not overcome the drawbacks of the prior art. Other references, such as the reference entitled “A 1-Gb/s Joint Equalizer and Trellis Decoder for 1000Base-T Gigabit Ethernet” written by Haratsch and Azadet and published in IEEE Journal of Solid State Circuit, vol. 36, No 3, March 2001 proposes implementations of equalizers and decoders based on the Rao reference. These implementations incorporate undesirable levels of complexity and thus may not be possible to implement.
In contrast to non-precoded systems, precoded systems perform some signal processing on the signal at the transmitter prior to transmission to account for inter-symbol interference. Such prior art systems suffer from an undesirable increase in transmit power and require greater dynamic range that offsets any gain achieved by preceding. In response to the drawbacks of other types of preceding, Tomlinson-Harashima (TH) preceding was proposed. The article “Comparison of Precoding Schemes for Digital Subscriber Lines” written by R. Fischer and J. Huber and published in IEEE Transactions on Communications, vol. 45, no. 3 March, 1997 discusses TH preceding. TH precoding implements a non-linear modulo operation to maintain the transmitted signal within a predefined range. This proposed solution undesirably requires a much larger dynamic range at the receiver to handle an extended symbol constellation. In addition, with TH Precoding most of the reduction in transmit power provided by constellation shaping is lost.
Another proposed solution was presented by Agazzi et. al (“10 Gb/s PMD Using PAM-5 Trellis Coded Modulation”) in a Broadcom Corp. proposal to IEEE 802.3ae, in Albuquerque, N. Mex. in Mar. 6–10, 2000. However, this proposed solution increases the complexity of the transmit precoder at high data rates.
In addition, these above-described systems do not address the degradation of coding gain from noise correlation created by a channel shortening filter and forward equalizer. The expected coding gains from the Viterbi trellis decoder are based on the assumption of additive white gaussian noise (AWGN). When the noise is correlated (or not white) then the expected coding gain is not realized.
As a result, there is a need for a method and apparatus to facilitate high-speed data transmission while maintaining low error rates that is capable of overcoming the drawbacks of the art.
A communication system for transmitting data over two or more channels is disclosed that overcomes the drawbacks of the prior art. In one embodiment the communication system comprises a de-multiplexer configured to receive data from a data source on a first number of conductors and output the data on a second number of conductors such that the second number of conductors is greater than the first number of conductors. Also included in this embodiment is one or more error correction encoders configured to receive the data on the second number of conductors and perform encoding on the data to create encoded data. A multiplexer is also provided to process the encoded data from the one or more error correction encoders on the second number of conductors and output the encoded data on a third number of outputs. There also exist one or more mapping modules configured to receive the encoded data from the multiplexer and generate symbol values representative of the encoded data. The symbol values may be selected from a constellation of available symbol values.
In one configuration of this embodiment the system further comprises one or more precoders configured to receive the symbol values and process the symbol values to at least partially compensate for the effects of transmission through a channel. In another configuration the system further includes two or more transmit processing modules configured to transmit the symbol values through two or more transmission lines. The one or more mapping modules may comprise mapping modules configured to perform PAM10 mapping on the data. It is contemplated that the constellation of available symbol values may comprise a constellation of 10,000 data points, 8192 of which are reserved to represent data. Regarding the error correction encoders, 8-state Ungerboeck trellis encoders may be utilized. As one advantage to this system, the one or more mapping modules may map the data into symbol values selected from a constellation of symbol values that have an average level of zero.
In another embodiment a communication system is disclosed for processing at least one signal received over two or more channels. This embodiment comprises one or more equalizers configured to receive one or more signals from the two or more channels and process the one or more signals to reduce the effects of transmission through the two or more channels. The communication system includes a de-multiplexer configured to receive the signal from the one or more equalizers on a first number of inputs and output at least two signals on a second number of outputs. It is contemplated that the second number of outputs is greater than the first number of outputs. Two or more error correction decoders are also included and configured to perform decoding on the two signals thereby creating at least two decoded signal. Further, the system includes a multiplexer configured to combine the at least two decoded signals into at least one decoded signal. The decoded signal is provided to one or more demapping modules configured to translate the decoded signal to data. This is but one possible example embodiment.
In one configuration of this system the de-multiplexer converts four signals into sixteen signals and the multiplexer converts sixteen signals into four signals and the two or more channels comprises four channels and each channel comprises a twisted pair conductor. It is contemplated that the error correction decoders may be configured to decode 8-state Ungerboeck trellis encoded data. The demapping modules may be configured to map 4D PAM10 symbols into data and control symbols and in one embodiment the signal comprises PAM10 symbols. It is further contemplated that the equalizers may comprise a feedforward equalizer and a feedback equalizer.
Also disclosed herein is a method of processing data prior to transmission over one or more channels to increase the data transmission rate. This method comprises receiving a first data stream from a data source and demultiplexing the first data stream into two or more second data streams. Thereafter, the method performs error correction encoding on the two or more second data streams to create two or more encoded data streams and it then multiplexes the two or more encoded data streams to create interleaved encoded data streams. The number of interleaved encoded data streams may be less than the number of non-interleaved encoded data streams. The interleaved encoded data streams may then be modulated to convert the interleaved encoded data streams into symbol values on two or more modulator outputs.
In one configuration the error correction encoding comprises trellis encoding and the step of modulating comprises modulating the one or more encoded data streams using ten level pulse amplitude modulation. The interleaved encoded data stream may be comprised of control codes and data and the step of modulating may translate the data and control codes into symbol values. In one embodiment the symbol values that represent data have an average level of zero. The step of modulating may comprise performing ten value pulse amplitude modulation.
In yet another embodiment a method for processing signals received at a first data rate is provided that comprises the steps of receiving two or more signals via two or more channels and filtering the two or more signals to reduce the effects of intersymbol interference. Thereafter, the method may de-interleave the two or more signals to create three or more signals at a second data rate and then perform error correction processing on the three or more signals to correct errors arising during transmission. Next, the method may interleave the three or more signals to create two or more signals at a third data rate and demodulate the two or more signals to convert the signals to data.
In one variation, the second data rate is lower than the first data rate and the third data rate is higher than the second data rate. The signal may comprise symbol values and the step of demodulating may comprise converting at least one symbol value to data. The step of demodulating may also comprise mapping the two or more signals to data and control codes. It is contemplated that the signal may comprise a symbol value and the demodulating may comprise comparing the signals to a constellation of symbol values to determine if the signals comprise data or a control code. In one embodiment the signal comprises ten level pulse amplitude modulated symbol values.
As a further advantage over the prior art, a method and apparatus for constellation mapping is disclosed that overcomes drawbacks of the prior art. The method and apparatus for constellation mapping disclosed herein provides a constellation, or group, of symbol values that provides a desirable average transmit power level. In embodiment the average value of the symbol values that comprise the available symbols to represent data, control codes, or both is zero. In this manner, use of an additional transmit polarity scrambler may be avoided.
One exemplary system for mapping data and control codes to symbol values operates by receiving a signal and analyzing the signal to determine if the signal comprises a control code or data. In this embodiment the average level of the symbol values reserved for data comprises zero. As a result of the analyzing, this method translates the data to a data symbol value such that the data symbol values are selected from a group of symbol values reserved for data. Also resulting from the analyzing, the method translates control codes to a control code symbol value. The control code symbol values are selected from a group of symbol values reserved for control codes
It is contemplated that in one embodiment the two or more inputs comprise four inputs and that the group of symbol values reserved for data comprises 8192 symbol values. In one embodiment the signal comprises binary information. In one exemplary method of operation, the analyzing comprises comparing the signal to a group of signal values reserved for control codes.
In another embodiment, a method is provided for creating a constellation of available symbol values available for mapping data into symbols comprising the steps of determining a total number of data values to be mapped into the constellation and defining a set of symbol values from the constellation reserved for data. This method also includes defining an association between data and a symbol value, such association may be used by a mapper. The average value of the set of symbol values reserved for data may comprise zero.
In one variation this method further comprising defining a set of symbol values from the constellation reserved for control codes. This constellation may comprise 10000 data points and each symbol value may comprise one of ten different levels. In one embodiment the constellation comprises a four dimensional constellation.
Yet another system for mapping data and control codes into symbol values may be provided such that in one embodiment an input is configured to receive data and control codes. The input connects to control logic configured to translate the data and control codes to a symbol value. The control logic may be configured to map a control code to a symbol value reserved for control codes and map data to a symbol value reserved for data. The symbol values reserved for data may be selected to have an average power level of zero.
A system for mapping data is also disclosed that comprises an input configured to receive a signal with a processor connected to the input. The processor is configured to analyze the signal to identify data and control codes, translate the data into symbol values reserved for data, and translate the control codes into symbol values reserved for control codes, such that the symbol values reserved for data have an average level of zero.
Such a system may further comprise a memory accessible by the processor and configured to store symbol values. The symbol values reserved for data may comprise 8192 symbol values. The step of translating the data into a symbol value may comprise translating two or more bits of binary data into a symbol value selected from a group of symbol values reserved for data and having an average value of zero. As discussed below, the symbol values may comprise ten level pulse amplitude modulated symbols.
Upon receipt of the symbol values a method of processing the symbol values may occur. In one embodiment the method comprises receiving a symbol value and analyzing the symbol value. Responsive to the analyzing, a system may convert the symbol value to either two or more bits of data associated with the symbol or a control code associated with the symbol. The group of symbol values that may be associated with the two or more bits of data may have an average value of zero.
In one embodiment the step of analyzing comprises comparing the symbol value to a plurality of symbol values stored in a memory to thereby map the symbol value to either two or more bits of data or a control code. It is contemplated that the step of receiving comprises receiving a symbol value during each clock cycle on each of four channels.
Other systems, methods, features and advantages of the invention will be or will become apparent to one with skill in the art upon examination of the following figures and detailed description. It is intended that all such additional systems, methods, features and advantages be included within this description, be within the scope of the invention, and be protected by the accompanying claims.
The components in the figures are not necessarily to scale, emphasis instead being placed upon illustrating the principles of the invention. In the figures, like reference numerals designate corresponding parts throughout the different views.
In general, the method and apparatus disclosed herein performs data transmission using time interleaved error correction coding in a multi-pair or multi-channel system to achieve data transmission rates and error rates that are better than that achievable by systems of the prior art. In reference to
The receive module 138 and transmit module 142 communicate with a processor 146. The processor 146 may include or communicate with memory 150. The memory 150 may comprise one or more of the following types of memory: RAM, ROM, hard disk drive, flash memory, or EPROM. The processor 146 may be configured to perform one or more calculations or any form of signal analysis. In one embodiment the processor 146 is configured to execute machine readable code stored on the memory 150. The processor 146 may perform additional signal processing tasks as described below.
The second transceiver 134 is configured similarly to the first transceiver 130. The second transceiver 134 comprises an interface 152 connected to a receiver module 156 and a transmitter module 160. The receiver module 156 and a transmitter module 160 communicate with a processor 164, which in turn connects to a memory 168. Operation occurs as described below in more detail.
The communication system illustrated in
The output of the de-multiplexer 304 connects to error correction coding (ECC) modules 308A–308M. The ECC modules 308A–308M may perform any type error correction coding as contemplated by one of ordinary skill in the art. In one embodiment the error correction coding performed by modules 308A–308M comprise 8-state Ungerboeck trellis coded modulation. In one embodiment the 8-state Ungerboeck trellis coded modulation is performed on the four interleaved data streams to create a four dimensional (4D) 8-state Ungerboeck trellis coded data stream. The particulars of 4D 8 state Ungerboeck trellis coded modulation are understood by one of ordinary skill in the art and hence are not described in detail herein. In other embodiments the following types of error correction encoding may be performed: trellis coding, convolutional coding, block coding, product coding or any other type of coding.
As an advantage of the system shown in
The outputs of the ECC modules 308A–308M connect to a multiplexer 312 configured to combine the two or more input signals into a reduced number of outputs. It is contemplated that the inputs to the multiplexer 312 may operate at a lower data rate than the output of the multiplexer.
The output of the de-multiplexer 512 feeds into processing modules 520. The processing modules 520 may comprise any combination of hardware, software, or both configured to performed any type processing as desired. In one embodiment the processing comprises error correction coding, such as 8 state Ungerboeck trellis coded modulation. Similar error correction de-coding may occur at a receiver after de-interleaving. As an advantage to de-interleaving, the data rate is reduced thereby providing more time between subsequently arriving bits or symbols at the processing modules 520. As a result, the processing modules 520 may operate at a lower speed, or may operate at the highest speed possible and operate in parallel to achieve higher overall transceiver operation.
The processing modules 520 provide the processed data to a multiplexer 530. The multiplexer 530 is configured to manipulate the processed data to a reduced number of outputs as shown. Internal operation of de-multiplexer 512 and the multiplexer 530 are known by one of ordinary skill in the art and accordingly are not discussed in detail herein.
Returning now to
The output of the mapper 320 connects to precoding modules 324A–324N. Although in one embodiment the mapper 320 includes four outputs, each of which feed into a precode module 324, it is contemplated that the precoding operations may be distributed among any number of precode modules, up to N modules, where N is a positive whole number. Precoding may comprise any type signal processing that may occur in the transmitter to counter the effects of the channel. In one embodiment the preceding comprises processing by a finite impulse response filter. Precoding may occur as described a co-pending patent application, entitled Method and Apparatus for Channel Equalization, by inventor William Jones, which was filed on Jul. 1, 2002.
The output of the precoders 324A–324N connect to one or more transmit processing systems 330A–330N. Any number of transmit processing systems 330A–330N may be utilized where N is any positive whole number. Transmit processing comprises processing to facilitate transmission of the signal over one or more channels. The transmit processing systems 330A–330N may comprise line interfaces, amplifiers, transformers, filtering, shaping filters, hybrids, digital to analog converters, and the like as may be desired to facilitate transmission of a signal or data to a remote location.
As an advantage to the configuration shown in
A convolutional encoder 630 connects to output O11 and output O12 of the scrambler 608 and is configured to provide an input to the mapping module 624. In the exemplary embodiment shown in
The output of the scrambler 608 and the output of the convolutional encoder become a thirteen bit encoder output word that is fed into the mapper 624. In the embodiment shown in
A data/control code selector input 660 connects to the scrambler 608 and the mapper 624 and carries control information to indicate the type input to the scrambler 608. As an advantage over the prior art, the method and apparatus described herein encodes the control codes as well as the data thereby providing the benefit of encoding to the control codes. To this end, the selector input 660 may be configured to provide the control codes to the scrambler 608 over an input 662. Inputs 664, 668 which also connect to the selector input 660 may provide notice to the scrambler and the mapper 624 that a control code is be provided to the scrambler. In one embodiment the inputs 664, 668 indicate a control code by providing a logical one value to the scrambler 608 and the mapper 624. It is also contemplated that during insertion of a control word, the scrambler may thus disregard the TX data input 604.
The output of the line interfaces 708 connect to an equalizer module 712. The equalizer module 712 comprises hardware, software, or both configured to perform equalization on the received signal. In one configuration the equalization occurs to reduce intersymbol interference. In one configuration the equalizer module comprises a feedforward filter and decision feedback equalizer. Co-pending patent application, entitled Method and Apparatus for Channel Equalization, by inventor William Jones, which was filed on Jul. 1, 2002, is incorporated in its entirety herein and describes an equalization system as is compatible for use with one or more embodiments of the invention. It is contemplated that any equalizer system or other apparatus configured to generate an accurate decision of the transmitted signal may be utilized. It is contemplated that the equalization module 712 may include a slicer or other decision device and be located or distributed at any point in the receiver.
The output of the equalizer 712 feeds into a de-multiplexer 716. As discussed above, the de-multiplexer 716 is configured to reduce the data rate of a signal by spreading or distributing the signal onto two or more outputs. Hence, conductor 720 comprises a greater number of conductors than is associated with the outputs 718. Hence, output 718 is divided or de-interleaved into conductor 720. The data rate on each conductor in input 720 may be less than the data rate on conductor 718.
The outputs 720 of the de-multiplexer 716 feed into an error correction decoding module (ECdC) 728. In the module 728 there may exist several ECdC sub-modules 732 configured in association with each of the outputs 720 as shown. This provides the advantage of distributing the processing burden of the error correction decoding across numerous subsystems. As a result, high-speed channel throughput may be achieved in a realizable architecture. As shown, the receiver may be configured with M number of ECdC modules 728 M where M is any positive whole number. Thus, the number of ECdC modules 728 is based on the degree of interleaving not the number of physical channels. Any level of expansion is contemplated.
It is contemplated that the error correction decoding is matched to that utilized in an associated transmitter. In one embodiment the ECdC modules 732 are configured to perform continuous trellis decoding. This is in contrast to packet based trellis coding/decoding. As a result, the benefits of continuous trellis decoding are gained. These benefits include but are not limited to a reduction in the total number of symbols that must be sent to fully error correction decode the received signal. In contrast to packet based decoding which requires that additional symbols be sent to terminate the trellis, the continuous trellis decoder embodiment operates on a continuous stream of trellis encoded data and control symbols and does not require that the trellis be terminated at the end of each packet. Continuous trellis decoding is discussed below in more detail.
The ECdC modules 728 output the error corrected signal or data to a multiplexer 734. In one embodiment the multiplexer 734 reverses the effect of the de-multiplexer 716 to assemble the data from the ECdC 728 into a reduced number of conductors. This interleaving process is described above in detail and accordingly is not discussed again. The output of the multiplexer 734 connects to a demapping system 740 which is configured to reverse the effects of mapping that occurred during the transmit processing.
In one embodiment the demapping system 740 reverses the effects of 4-D PAM10 mapping.
In one embodiment constellation shaping occurs on the data during the mapping process. In general, mapping may occur on the data to realize an increase in the effective data rate. In one embodiment the mapping that occurs is PAM10 (ten level pulse amplitude modulation).
Using a PAM10 modulation or coding scheme the following one dimensional (1D) constellation may be generated:
It is contemplated that any number of bits may be transmitted per baud. In one embodiment 12 bits per of data are combined with an extra bit generated as a result of trellis encoding for a total of 13 bits per baud being transmitted. To transmit the 13 bits per baud requires 8192 constellation points. Subtracting the 8192 required constellation points from the 10,000 total constellation points that are available from 4D PAM10 mapping provides 1808 remaining constellation points not used for data transmission.
In one embodiment of mapping the unused 1808 constellation points are selected for use as control symbols. A control symbol is defined herein to mean any code or output value combination that, in contrast to data, serves as an indicator for system operation. In one embodiment the control symbols are sent by the transmitter and detected by the receiver to control or assist in the transmitter and receiver operation. The control codes may signify any of the following control activities, Start of Packet delimiter (SOP), End of Packet delimiter (EOP), Error (E), Idle (I), Remote Fault (RF), Sequence Ordered Set (SOS) or others as needed.
It is contemplated that any unused symbols, i.e. symbols not designated by the mapper to represent one or more bits of data, may be used to represent one or more control symbols. Selection of which constellation points are available for data and which are available for control symbols may occur to establish the average level of the symbol values reserved for data at zero.
In one embodiment the control symbols are selected from the group of constellations that contain at least two −9 or +9 values or a −9 value and a +9 value. In other embodiments other constellation values may be selected.
Selection of constellations that contain at least two −9 or +9 values or a −9 value and a +9 value in their constellation for use as control symbol identifiers has many advantages over prior art constellation assignment schemes. One advantage is that the selection of these values shape the constellation used for data transmission. As a result of such shaping, transmit level over time is symmetric with respect to a zero level since there are an equal number of constellation points used for data transmission that are greater than zero and less than zero. Thus, there is an equal distribution for symmetric constellation points. In particular, there are 896 data points for each of the values −7, −5, −3, −1, 1, 3, 5, and 7. The 1-dimensional values +9 and −9 have a distribution in a 4D constellation of 512 occurrences each. Thus, removing the constellations having at least two of either or both of −9 or +9 values from those available to represent data provides a symmetric level distribution during data transmission.
The even distribution achieved by such shaping provides the advantage that a system adopting this constellation shaping method will not require an additional data transmit polarity scrambler at the output of a mapper system for the purpose of generating a symmetric transmit signal on average. An additional scrambler is not required because over time the evenly shaped constellation selects level values during the mapping process that evenly distributes the output signal to have a zero level over time. Hence, an additional scrambler component may be eliminated from such a system thereby resulting in lower complexity.
A further advantage is that transmission of symbols may occur with a lower average power since 4D constellations that include two −9 values, two +9 value or both a −9 value and a +9 value are not available during data mapping. While the peak power transmitted remains at the highest mapping value, i.e., +or −9 level in the example embodiment described herein, the overall average power transmitted over time is reduced. This provides the advantage of increasing the distance between transmit levels given a limit on average transmit power thereby leading to an effective increase in signal-to-noise ratio at the receiver.
Yet another advantage is that it allows for continuous trellis coding/decoding where the control symbols may be processed through the error correction coding and hence realize the coding gain resulting from error correction coding. In systems of the prior art, control symbols were not processed through the error correction coding and decoding. As a result, systems of the prior art were more likely to miss an important control code which would disrupt system operation. In one embodiment the control symbols are subject to trellis coding and decoding and thereby receive an approximately 6 dB coding gain.
It is contemplated that in other embodiment the constellation may not be balanced about zero but instead has an average power level of no more than 15% of the maximum constellation value. In such an embodiment the constellation is still close to zero thereby gaining some of the benefits described above. In such an embodiment an additional scrambler may be required or more than mapping constellation may be used to obtain an average transmit level over time of zero.
In reference to
Thereafter, at a step 916 the transmit system may perform error correction coding on the data on each of the fractional rate data streams. In one embodiment the error correction coding comprises 8-state Ungerboeck trellis coded modulation. In other embodiments trellis coding, convolutional encoding, block coding, or product coding may be performed. In one embodiment the error correction coding generates a 13 bit word that is output from the error correction encoder at a rate of 10/48 GHz.
After encoding, the data is provided as an input into a multiplexer. At this stage the data may be referred to as one or more fractional rate encoded data streams. The multiplexer selectively interleaves or arranges the symbols into one or more full rate encoded data streams. This occurs at a step 920. In one embodiment the multiplexer receives a first number of data streams and as a result of the interleaving outputs a second number data streams, wherein the first number is greater than the second number. In one embodiment the output from multiplexer is 13-bits wide and operates at a data rate of 10/12 GHz.
At a step 928 a mapper performs mapping on the encoded data. While any form of mapping may occur, in one embodiment 4-dimensional (4D) pulse amplitude modulation type mapping (PAM) occurs on the full rate encoded data. In one embodiment ten level PAM mapping may occur. At a step 932, the mapped data is output from the one or more mapping modules as symbol values. As is contemplated with multidimensional mapping, each mapping module includes two or more outputs. In one embodiment configured with 4D mapping each mapping module generates 4 outputs. In one embodiment each mapper receives a single multi-bit wide input during a clock cycle and outputs four symbols during a clock cycle. In one embodiment the outputs operate at a data rate of 10/12 GHz. Certain aspects of mapper operation are discussed below in more detail on conjunction with
At a step 936 the transmitter may perform preceding. Precoding may be performed to reduce the unwanted effects of intersymbol interference on the signal. At a step 940 additional transmit processing may occur on the data.
Thereafter, at a step 944 the data, represented as symbols, is transmitted over the two or more paths. Although described herein in terms of twisted pair conductors and channels, it is contemplated that any medium may be utilized, including but not limited to, optical, and wireless, such as radio or other frequency. It is contemplated that a path, such as a twisted copper pair channel, corresponds to each channel over which data is transmitted. In one embodiment, the transmit rate on each of the channel is 10/12 GHz. In one embodiment, there exist four transmission lines, each of which comprises a pair of twisted copper conductors. An example of such a line is category 5 UTP cabling compatible with the Ethernet 1000BASE-T standard. In other embodiments any other communication channel may be utilized, including but not limited to one or more, wireless channels, any electrically conductive channels, fiber optics or fiber optic cables, optical channels, twisted pair conductors, coaxial cabling, or any other type of channels as may be contemplated by one of ordinary skill in the art.
As an advantage of the method and apparatus described herein, the number of available constellation points is greater than the number of possible 4D symbol combinations used to represent data that must be assigned constellation points. In one embodiment the extra available constellation points are available and utilized to represent control codes.
Turning now to
In one embodiment the constellation mapping scheme for data values excludes from use those 4D constellations reserved for control codes. For example, a PAM10 constellation may comprise the values −9, −7, −5, −3, −1, 1, 3, 5, 7, and 9. In one embodiment step 1004 excludes from the set of constellation points available for representing data the constellations containing two or more ±9 values. It is contemplated that the mapping scheme is known by the receiver and the transmitter and may be hardwired or programmed during an initialization phase or during manufacture.
At a step 1008 the mapper is provided an assignment scheme for mapping of control codes. In one embodiment the control code assignment scheme assigns control codes to four dimensional constellation values that include two or more ±9 values. Thus, these four dimensional constellations are reserved for control codes.
At a step 1012 the mapper receives an input. In one embodiment the input comprises a 13 bit word that is mapped into a four dimensional PAM10 constellation. In such an embodiment the mapper may include four outputs each of which carry a PAM10 value during an output event, such as a clock cycle.
After receiving the input, the mapping module determines if the received input (binary data) comprises data or a control code. This occurs at a step 1016. In one embodiment a look-up operation occurs during the mapping processes. In another embodiment an algorithmic operation is executed as part of the mapping or demapping process. In one embodiment the control code arrives at the mapper as two or more binary bits.
If the input comprises a control code then the operation advances to a step 1022 wherein the mapper maps the control code into a four dimensional constellation comprised of constellation points or symbol values that are reserved for control codes. In one embodiment this comprises constellation points that contain two or more ±9 values. After mapping, the mapper operation advances to a step 1026 and outputs the four dimensional constellation as symbols on four outputs. It is contemplated that the receiver be provided with the mapping scheme so that upon reception and demapping of a constellation, the receiver is thus able to determined that the received constellation is a control code. At the receiver appropriate interpretation and processing of the control code may occur.
Alternatively, if at step 1016 the input is interpreted to be data then the operation advances to a step 1030 and the mapper maps the data into an appropriate constellation. In one embodiment this may comprise a look-up operation. In one embodiment the mapping comprises mapping the data into a constellation selected from a group of constellations that have an average level of zero. In such an embodiment the constellation values have been selected or shaped to have an average level or value of zero thereby eliminating the need for a second scrambler on the channel side of the mapper. Further, a lower overall transmit power may result of the maximum constellation values being reserved for the less often occurring control codes.
In one embodiment a four dimensional constellation is selected that does not contain two or more ±9 values. It is contemplated that these constellations are reserved for control symbols. Next, at a step 1034 the mapper outputs the four dimensional constellation of data symbols on four outputs. It should be noted that a four dimensional constellation size is selected for purposes of discussion and not limitation. Any size or dimension of constellation and any type of modulation or mapping may be selected.
An advantage to the method discussed in
After equalization the signals are provided to a de-multiplexer. The de-multiplexer separates the signals into fractional rate symbol stream. This occurs at a step 1112. As a result the data rate in each data stream is less than prior to de-multiplexing. This facilitates parallel processing of high rate data streams. In one embodiment the multiplexer outputs sixteen signals each of which comprise a symbol value, such as a PAM10 signal at a data rate of 10/48 GHz.
At a step 1116 the receiver performs error correction decoding on the signals output from the multiplexer to create one or more fractional rate decoded symbol streams. As is understood by one of ordinary skill in the art, the type of error correction decoding is linked to the type of error correction encoding performed at the transmitter.
At a step 1120 the system multiplexes, i.e. interleaves, the symbol streams to a one or more full rate symbol streams. At a step 1124 demapping is performed on the symbols. The demapping reverses the effects of mapping performed at a transmitter. In one embodiment, not shown, the receiver includes four demapping modules each of which are configured to receive four inputs and output a single output. In one embodiment the output is 12 bits wide and operates at a rate of 10/12 GHz. In one embodiment the output is 13 bits wide and operates at a rate of 10/12 GHz. As discussed in conjunction with
In the implementation example discussed herein, the selection of FEC is a tradeoff of implementation complexity versus the desired coding gain. At a baud rate of 10/12 GHz it is desirable to obtain a low complexity implementation. As described above, the 4D 8-state Ungerboeck trellis code is selected as a baseline code. When presented with additive white gaussian noise (AWGN), this FEC code on 10 PAM achieves a coding gain of about 3.5 dB over PAM8. In addition, shaping gain of about 0.64 dB may be realized by using the described constellation shaping.
The 4D 8-state Ungerboeck code is sensitive to error propagation from the DFE and noise correlation created by the forward equalization and channel shortening equalization. In the implementation example discussed herein 4D coding across pairs was selected and the trellis coding was 4-way interleaved across time. This reduced the complexity of the design so that 4 trellis decoders operating at ¼ the clock rate were utilized. In addition, the 4-way time interleaving of the trellis code prevents bursts of error from hampering trellis code operation. A burst of errors is thus spread out across four time interleaves of trellis coding and subject to FEC decoding in different decoders.
In the example implementation of
The 8-state Ungerboeck trellis encoder 1208 outputs a thirteen bit wide word at a rate of 10/48 GHz into a multiplexer 1212. The 4 sets of inputs to the multiplexer 1212 at 10/48 GHz are interleaved to become a full rate encoded data stream comprising 13 bits at 10/12 GHz. The output of the multiplexer 1212 feeds into a 4-D PAM10 mapper 1216. The output of the mapper 1216 comprises one or more full rate modulated symbol streams.
As can be understood, the selection of the number of PAM levels, the baud rate, and the forward error correction (FEC) coding are all interrelated. In the implementation example described herein the channel capacity is maximized by using a baud rate of around 800 MHz.
The implementation example provided herein also includes constellation shaping. Given a selection of four pairs with PAM10 line decoding, the 4D constellation has 10×10×10×10=10,000 constellation points. The twelve bit data word has 4096 distinct values and the thirteen-bit trellis encoded word has 8192 distinct points. The PAM10 mapping results in the thirteen-bit word being mapped into one of 8192 possible constellation points. Therefore, there exist 1808 redundant points, some of which are used for control symbols.
The implementation example obtains some shaping gain by eliminating the constellation points with the largest transmit power. In this example, the +9 and −9 symbols have the largest transmit power in any single dimension and there are exactly 1808 4D constellation points that contain any combination of two or more +9's and/or −9's. Therefore, the shaping eliminates from the normal data mapping any point that contains one of these combinations. The remaining points are uniquely mapped to the remaining constellation points using any manner known in the art. In one embodiment an algorithmic equation is utilized during the mapping process. In this embodiment the rate of occurrence for a +9 or −9 on any given pair is 4/7 the rate of any other levels.
Precoding and transmit processing 1220, 1224 may occur as known in the art to facilitate transmission over four twisted pair conductors. In this implementation example, transmission of the PAM10 symbols occurs over the four twisted pair conductors at a rate of 10/12 GHz.
While various embodiments and implementation examples of the invention have been described, it will be apparent to those of ordinary skill in the art that many more embodiments and implementations are possible that are within the scope of this invention.